FogQN: An Analytic Model for Fog/Cloud Computing

Several tradeoffs need to be considered when determining the optimal fraction f of data processing executed at the cloud versus at fog servers. The processing capacity of fog servers is typically smaller than that of cloud servers. On the other hand, it may be more expensive to use cloud resources as opposed to fog servers. As f increases, more data has to be sent and received from the cloud. On the other hand, if too much processing is left for the fog servers, they may not have enough capacity to handle requests from sensors and other IoT devices and may become a bottleneck. This paper presents an analytic model and a publicly available tool, called FogQN, based on open multi-class Queuing Networks (QN) for fog and cloud computing. FogQN was validated with the JMT simulation tool using both distribution-based arrival rates and inputs from real IoT applications.